AI Pipeline Forecasting
AI pipeline forecasting estimates how much construction revenue may convert from active pursuits. It uses historical outcomes to score each opportunity's win probability and timing alongside the BD team's judgment.
Why it matters in construction
A GC’s revenue forecast is backlog plus whatever the pipeline converts. Backlog is known. The pipeline is where the forecast goes wrong, and the errors compound. A pursuit marked 70 percent that was really a 30 percent shot inflates revenue, which inflates the staffing plan, which drives a PM hire the firm did not need.
Many firms use a Friday pipeline meeting to revisit those numbers. AI pipeline forecasting adds evidence from similar past pursuits. It can compare the current job with past negotiated healthcare work of a similar size, owner type, and level of precon involvement, then use those outcomes to weight the forecast.
How it works
- Describe the pursuit. Each opportunity carries its delivery method, project type, size, owner relationship history, whether the firm was invited or found it, competitor count, and how much precon work has already gone in.
- Learn from outcomes. Past pursuits with recorded win/loss, award date versus expected date, and final contract value train a model to estimate win probability and timing slip for each profile.
- Read the activity. A language model scans CRM notes, meeting summaries, and email for signals not captured in the structured fields, such as an owner asking for a GMP, an architect raising a budget concern, or several unanswered follow-ups.
- Weight and roll up. Each pursuit gets a calibrated probability and an expected start month. The roll-up is weighted revenue by month with a range, stacked on top of backlog.
- Flag drift. Pursuits whose stated probability differs sharply from the model’s estimate go on a list for a conversation.
The output needs to explain the probability. Otherwise, the team has no basis for judging whether the software’s estimate is more useful than the BD lead’s.
Example in practice
In one hypothetical scenario, a commercial GC has $85M in active pursuits across 18 opportunities. The CRM says weighted value is $41M. The model says $29M.
Three pursuits explain the gap. A $22M public bid marked 50 percent scores at 18 percent, because the firm’s hard-bid win rate at that size against six or more bidders is under 20 percent. A $15M negotiated office job marked 40 percent moves up to 65 percent, because the owner awarded the last two projects to the firm and precon has been billing for four months. And a $12M pursuit shows 45 days of no logged activity, which in this firm’s history precedes a loss more often than not. Leadership pulls the $12M from the staffing plan, holds the PM hire tied to the public bid, and moves the office job into the workforce forecast.
Frequently asked questions
Why not just use the probability the BD lead enters in the CRM?
You can, and most firms do. Those numbers are often not revisited and can be optimistic. AI pipeline forecasting compares them with outcomes from similar pursuits and can adjust the forecast.
How much history does the model need?
Enough won and lost pursuits to see patterns, typically three to five years of CRM data with outcomes recorded. If your CRM only has wins, the model has nothing to learn from.
Does it forecast timing or just probability?
Both. Start-date slip is often a bigger forecasting error than win rate. Owners delay awards, permits lag, and a job you win in March may not bill until August.
Go deeper
- From the blog Revenue Forecasting for General Contractors: A Practical Framework Most GCs track project financials but can't answer what they'll bill next quarter. Here's the three-layer framework that connects WIP, backlog, and weighted pipeline into a rolling revenue forecast.
- From the blog Business Development for General Contractors: 4 Steps Business development for general contractors belongs to preconstruction. Here's a four-step framework for running BD on one pipeline with estimating.
- From the blog The Financial Metrics Your Preconstruction Team Should Be Tracking (But Probably Isn't) Most GCs track project financials but ignore preconstruction metrics. Here are 5 KPIs that measure whether your estimating effort is actually paying off.